Analog Fundamentals of the ECG

Signal Chain
Prepared by
Matthew Hann,
Texas Instruments
Precision Analog Applications Manager
Presented by
Jose Duenas,
Precision Analog Product Marketing Engineer 1
 Objectives

• Introduce Basic ECG Concepts

• Motivate the Need for TINA and SPICE
Simulation for ECG Analysis
• Introduce Discrete Analog Functions of
the ECG Signal Chain
• Motivate Need for Low Cost Integrated
ECG Conditioning System
• Introduce the ADS1298 and Its
Embedded ECG Circuitry and Functions
2
 Analog Fundamentals of the
ECG Signal Chain
• What is a Biopotential?
• What is ECG?
• The Einthoven Triangle
• Analog Lead Definitions, Derivations, and Purpose
• Modeling the Electrode Interface
• Input Filtering and Defibrillation Protection
• The INA front end
• AC vs. DC coupling
• Right Leg Drive (RLD) Amplifier Selection and Design
• The ECG Shield Drive
• Lead Off Detection
• PACE Detection
• INA post Gain + Analog Filtering
• A/D Conversion Options and Filtering
• ADS129x Introduction, Features, and Advantages
3
What is a Biopotential?

41
 What is a Biopotential
An electric potential measured between living cells

5
 What is a Biopotential
Every cell is like a little battery

Cell membrane
(lipids)
Intracellular medium/
Cytoplasmic fluid

Extracellular •Resting ion chs.

medium/fluid Ion channels
•Active ion pumps
(proteins) •Gated ion chs.
6
 What is a Biopotential
Every cell is like a little battery

Cell membrane
(lipids)
K+ is 30 to 50x
higher @ rest

Electrostatic
Force

Na+ conc. is 10x

higher @ rest Ion channels •Resting ion chs. Diffusional
•Active ion pumps force
(proteins) 7
•Gated ion chs.
 What is a Biopotential
Every cell is like a little battery
Neurotransmitter are
received at dendrites
Cell membrane
(lipids)
Intracellular medium/
Cytoplasmic fluid

Extracellular •Resting ion chs.

medium/fluid Ion channels •Active ion pumps
(proteins) •Gated ion chs. 8
 What is a Biopotential
Biopotentials from cells electrodes

9
What is ECG?

101
 What is ECG?
A measure of the electrical activity of the heart

11
 What is ECG?
ECG and blood pressure waves

12
What is ECG?
Actual ECG-normal

13
 What is ECG?
ECG irregular tracings due to external artifacts

14
 What is ECG?
Modeling the electrode interface

Electrical characteristics include a DYNAMIC resistance,

capacitance, and offset voltage 15
Analog Lead Derivation

161
 Analog Lead Derivation
ECG Einthoven Triangle, 1907

3 Body Electrodes,
3 Derived Leads = I, II, III

LEAD I = VLA-RL – VRA-RL

LEAD II = VLL-RL – VRA-RL
LEAD III = VLL-RL – VLA-RL

Einthoven’s Law

Right Leg In electrocardiogram at any

Reference, given instant the potential of
RL any wave in Lead II is equal
to the sum of the potentials
in Lead I and III.
17
 Analog Lead Derivation
The Wilson Central (WCT) Provides Chest Lead
Reference at Center of Einthoven Triangle

Assuming:

RRA = RLA = RLL

Then:
WCT RA  LA  LL
3
RRA RRA

RA  LA  LL

WCT
3

18

*Drawing Taken From Bioelectromagnetism, Jaako Malmivuo and Robert Plonsey

 Analog Lead Derivation
The Wilson Central is the AVERAGE potential
between RA, LA, and LL

RRA

RLA
VOUT = V(1-6) - VWCT

RLL The Wilson Central

is used to derive a
reference potential
for the Chest Leads,
V1 – V6
19
 Analog Lead Derivation
Chest Lead Signals Provide Different Information at
Different Cross-Sectional Angles

Different Chest Leads

Provide:
*Unique ECG Signature
*Enhanced Pattern
Recognition
*Isoelectric Point @ V3-V4

20
 Analog Lead Derivation
Augmented Leads Derived via WCT to Provide
Enhanced Vector Information

Each lead provides

unique information about
the ECG Output Signal
Multiple Angles Give a
Better Than 2-D Picture
of the ECG Output
AVR, AVL, AVF derived
via midpoint of 2 limbs
(resistor divider) with
Respect to 3rd limb

21
Analog Lead Derivation
IEC60601-2-51—Diagnostic
Standards Electrodes Needed

1 Lead LA, RA
3 Lead LA, RA, LL
6 Leads LA, RA, LL
12 Leads LA, RA, LL, V1-6

22
 Analog Lead Derivation
Different Lead Combinations Reveal Axis Deviation

QRS in LEAD I QRS in AVF Left Axis

Deviation
(LAD)

and

Mean QRS Vector Points Toward Area of Infarction (Damage) 23

 Analog Lead Derivation
Different Lead Combinations Reveal Axis Deviation

I I

D e
R em

LA
tr
AVF AVF
Ex
A

D
al
I
R

m
I
A

or
D

AVF N
AVF

24
ECG Input Filtering, Defibrillation
Protection, and Isolation

251
 ECG Input Filtering and Protection
Example: LEAD I Protection with Input Filtering

Series Resistance CF, Ccm, and Cdiff + Protection

Limits Input Current Rfilter form LPF Diodes

+Vs

RPatient Rfilter1 Rfilter2

LA
CF CCM +Vs
NE2H Zener
Cdiff Diode
Rfilter1 Rfilter2 -Vs
RPatient
RA
NE2H CF CCM

-Vs

Patient
Protection 10 Ne2H Lamps/TVS Protect Cdiff = 10 x Ccm
-20k ohms Against Defibrillation
Voltages
26
System Block Diagram

271
 System Block Diagram
Wilson INA MUX
I QRS and PACE
II dV/dt Detect
RA
III
LA
aVR

LL aVL

aVF

RL Filter MUX Isolation

WCT
INA A/D
V1
V2
DIO
V3

V4
V5

V6 28
The INA Front End

291
 The INA Front End
Key Features of the INA Front End

*Important* Less Important

Input Bias Current Input Offset Voltage

Input Impedance Input Offset Voltage Drift
Input Current Noise Gain Error
Input Voltage Noise Nonlinearity
Power Consumption PSRR
DC/AC CMRR

*DC Errors such as VOS are swamped out by the

Offsets Introduced by the Skin-Electrode Contacts

30
 The INA Front End
Ideal Simulation Circuit with Current and Voltage Noise
Sources

ECG + Skin Electrode Input CM + Ideal INA

Impedance Impedance + Differential Front End
Offset Filtering
C6 100p C7 47n
R19 1k

nV
V4 0 V n12 Vnoise
R21 1k R1 63.4k R3 63.4k

C12 33p C19 33p

R14 100k R20 1k

C10 33p C18 33p

Vout
+

C20 330p
ECGp + +
INA 1
C9 47n - -
Vref C8 100p
Vref
R23 1k V5 0 R4 63.4k R5 63.4k

R24 1k
R25 1k R22 100k fA
In12 In_fA

31
 The INA Front End
Simulation Showing Output-Referred Total RMS Noise
vs. Bandwidth (G = 1-10)
-Vdif Vin-
2+
Va1
+ Noise may NOT necessarily
- 150k 150k
A1
increase linearly with gain
50k
(INA or PGA topology
VCM
Rg 1k

- A3 Vout
+ +
dependent)
50k

- 150k 150k
T 32.00u
+ A2
+
Vdif Va2
Vin+ 28.00u
2

24.00u

Total noise (V) 20.00u

16.00u

12.00u

8.00u

4.00u

0.00
1.00 10.00 100.00 1.00k
Frequency (Hz)

32
 The INA Front End
TINA Simulations Showing Output-Referred ECG Signal
(G = 1-10)

Snapshot of R
Wave from
ECG Waveform

T 2.51
Voltage (V)

2.50

ECG Signal Varies

LINEARLY with
Increase in Gain
33
2.50
8.37m 366.11m 723.85m
Time (s)
 The INA Front End
What is the MAX gain on the INA When Using a DC
Removal Circuit?
Integrator Removes
Gained up DC offsets and (3)
servos INA output to Vref
Vref

(1) (2)

(1) Electrode Offset MAX = +/- 300mV

(2) Swing of INA = V(+) – 50mV
(3) Integrator Compliance = (ECGp + ECGn +VOS + VOSelectrode)* Gain < VCC - Vref 34
 The INA Front End
Simulation Circuit with Ideal INA and Vref = 2.5V as
Integrator Input
47n
1k
150m 63.4k 63.4k Vout
VCVS1 1
+

ECGp 52k + +

33p
33p

100k
- - Vref

330p
Vref

33p
+

47n
ECGn

33p

100n

+
-
-150m
VCC

+
63.4k 63.4k
1k OPA333
52k

OPA333 Used as Integrator

to Remove DC and Simulate
Real Response During
Saturation
35
 The INA Front End
ECG + Integrator Output of INA vs. Gain for Vref = 2.5V
2.50

Vout[G = 1]

2.50
2.50

Vout[G = 3]

2.50 As the output of the

2.50

Vout[G = 6]
Integrator Saturates from
Increased Gain, Vout of the
2.50
2.50 INA pulls away from Vref
Vout[G = 8]

2.49
2.71

Vout[G=9]

2.70
3.01

Vout[G=10]

3.00

0.00 750.00m 1.50

Time (s) 36
 The INA Front End
If it is Advantageous to Maximize Gain with a Low Noise
INA up Front, Why not AC Couple?
ECGp

VCM + ECGp +
VOSelectrode1

(ECGp –ECGn)*Gain

VCM + ECGn +
VOSelectrode2

ECGn

*AC Coupling Removes Electrode

Offsets so that Higher Gain Can Be
Used for Potentially Better SNR in the
Signal Chain Path
37
 The INA Front End
TINA Simulation Circuit to Show AC-Coupled INA Gain
Sweep

C2 47n R14 10M

Vref
V3 0 R8 63.4k R9 63.4k CCp 1u
R4 52k

C5 33p
C6 33p
+

INA 1
Vref ECGp + + +

C9 330p
Vout

C8 33p
- - -
+

C4 47n

C7 33p
ECGn
V4 0
R10 63.4k R11 63.4k CCn 1u Vref
R7 52k
Vref

R13 10M

38
 The INA Front End
INA Gain = 1-1000 with VREF = 2.5V AC Coupled
3.06

2.79

2.52
Voltage (V)

2.25

1.98

1.71

723.18m 969.60m 1.22

Time (s) 39
 The INA Front End
What is CMRR? Why is it Important in ECG?

ECGp

V out V OS Gain

VCM

ECGn Lower CMRR = More

Unwanted Output Signal

 V CM   V CM Gain 
CMRR 20log 10  20 
V OS V out
   
 V CM 
Vout  ECGp  ECGn  VOSelectrode  VOSOPA    Gain
  CMRR 
 20 
 10  40
 The INA Front End
What is CMRR? Why is it Important in ECG?

RP1
ECGp CC1
The INA includes the
R and C and must be
Amp considered in the
overall CMRR Analysis
VCM

ECGn CC2
RP2
•Mismatch in Rp and Cc
causes a differential
signal error

2 2
•Even a 1% tolerance
Rp1 Rp2 on Cc cause a 20dB
Vdiff_actual Vinp   Vinn 
2 2 attenuation in CMRR
Rp1  
j 
Rp2  
2 2 j 
 
 Cc1  Cc2
41
 The INA Front End
50/60Hz Common Mode Simulation Circuit with 1µF
Coupling Capacitors Mismatched

CMR TINA Circuit Test By

Sweep of Mismatch of Input
Coupling Capacitors

C2 47n R14 10M VCC Vref

Vref
V3 0 R8 63.4k R9 63.4k C1 1u
R4 52k C6 33p
+

C5 33p
V+
RG
C9 330p Vout

R1 20k
Ref
Out
C8 33p

C4 47n RG
+

V-

C7 33p
VCM V4 0
C12 1u
- U1 INA333_C
R10 63.4k R11 63.4k
R7 52k
Vref

R13 10M

42
 The INA Front End
Plot of CMRR vs. Frequency for .01 - .5% Coupling
Capacitor Mismatch

T -40.00

.5% Mismatch

-60.00
Voltage (V)

-80.00

0% Mismatch
-100.00
100.00m 1.00 10.00 100.00
Frequency (Hz)
43
 The INA Front End
Plot of ECG Response to 5Hz CM Input Signal (0%-
.5%) CC Mismatch
T 2.51

Lower Frequency
Signals Couple
Directly Into the ECG
Signal to the Output
Voltage (V)

2.50
0.00 400.00m 800.00m
Time (s) 44
 The INA Front End
Plot of ECG Response to 50/60 Hz CM Input Signal

T 2.51 •50/60Hz Noise Rejection is

Virtually Unaffected by AC
Coupling
•Dependent Primarily on the
Noise Magnitude and the
CMRR of the Front End INA
Vout

2.50
0.00 400.00m 800.00m
Time (s)

45
The Right Leg Drive Amplifier

461
 The RL Drive Amplifier
The RL Drive Amplifier Serves 2 Purposes: (1)
Common Mode Bias (2) Noise Cancellation

ECGp

CMnoise

CMnoise

ECGn

Average VCM is
*Tapping off center of split gain
Inverted and Fed resistor feeds the following
Back to RL; voltage to the RL Drive Circuit
Cancels 50/60Hz [(Vcm+ ECGP )+ (Vcm+ ECGn )]/ 2
noise = Vcm + (ECGp + ECGn )/2

47
 The RL Drive Amplifier
Simulation Circuit for Response to 50/60 Hz CM Noise
Injection Source

VCC
NS3 2.5
NS

Vref
R1 1k C2 47n +
+

U1 OPA333
R15 1M -

R4 1k VCC
R8 63.4k R9 63.4k
C10 1u
+

R12 1M
C5 33p
C6 33p
ECGp +
V+
RG

C9 330p
RG1 10k Ref Vout
Out
+

C8 33p
ECGn
+

RG
Noise

V-

C7 33p
RG2 10k - NS
U3 INA333
R10 63.4k R11 63.4k
NS2 2.5
C4 47n

R2 1k R7 1k

Vref
R5 1M = Included for
NS1 2.5
NS TINA spice
R6 1k R3 10k
- - Convergence
U2 OPA333 U4 OPA333
+ +
+ Vref +
R14 1M
R13 1M

R16 1M
VCC VCC

48
 The RL Drive Amplifier TINA
Simulation with NO RL Drive; CM Noise is Coupled to
Output
T 750.00u
ECGn

-150.00u
750.00u
ECGp

-150.00u
1.00
Noise

-1.00
2.51
Vout

2.50
0.00 665.73m 1.33
Time (s) 49
 The RL Drive Amplifier
TINA Simulation with RL Drive; Output Noise is
Reduced
T 750.00u
ECGn

-150.00u
750.00u
ECGp

-150.00u
1.00
Noise

-1.00
2.51
Vout

2.50
5.46m 282.16m 558.86m
50
Time (s)
 The RL Drive Amplifier
Analyzing the RLD Amplifier Loop

VECGp + Vep + VCM

  Vref  VoutA1  ECGp

VRL    RF  Vref Vep
 RI

RG/2 RA

RG/2
•More Gain = Better CMRR RA

•Loop Corrects for RF

Ven
Electrode Offset, VOSA1, ECGn
and VOSRLD
VREF
RI
VECGn + Ven + VCM

  Vcm  VECGp  Vep   Vcm  VECGn  Ven  RG 

VoutA1     2   VOSA1
 RG
  51
 The RL Drive Amplifier
Simulation Circuit for CMRR of RLD Loop

VCC
Sweep Feedback

Vref
Resistor Gain to
show Effects of +
+

CMRR @ 50/60Hz U1 OPA333

R15 1M -
C2 47n
VCC
R8 63.4k R9 63.4k
R4 52k C10 1u

R12 1M
V3 0

C5 33p
C6 33p
+
V+
RG

C9 330p
Vout
+

RG1 25k Ref

Vcm V4 0C4 47n Out

C8 33p

R17 1M
RG
V-

C7 33p
RG2 25k - U3 INA333
R7 52k
R_RLD 52k

R10 63.4k R11 63.4k

C_RLD 47n

VCC

NS1 2.5
R2 100k NS
RF 10k Vref V1 2.5
-
U4 OPA333
R3 10k +
+

R14 100k
-
V2 2.5

R16 10M
+
VCC
+
R13 100k

Vref
VCC
U2 OPA333
52
 The RL Drive Amplifier
CMRR Plots vs. Gain in RLD Loop
T -80.00

RF = 10k
CMRR (dB)

RF = 100k
-100.00
RF = 1M

RF = 10M

-120.00
10.00 100.00
Frequency (Hz)
53
 The RL Drive Amplifier

VCC
RL Drive Stability Simulation Circuit

Vref
Electrode +
+

Resistance Varies U1 OPA333

R15 1M -
With Contact and C2 47n
VCC
Moisture, Presents
R8 63.4k R9 63.4k
Problems for RLD V3 0 R4 52k C10 1u

R12 1M
Stability

C5 33p
C6 33p
+
V+
RG

C9 330p
RG1 25k Ref
C4 47n Out

C8 33p
V4 0

R17 1M
RG
V-

C7 33p
RG2 25k - U3 INA333
R7 52k
R_RLD 52k

R10 63.4k R11 63.4k

C_RLD 47n

VCC

R5 10M
Vref V1 2.5
R2 100k
L1 1G R3 10k -
U4 OPA333
- +
+

R14 100k
+
+ Vloop V2 2.5
Local RLD Loop is VoA VCC
Vref
Broken to Ensure
C1 1G

VCC
U2 OPA333
Proper Phase Margin
+

Over Range of Vin

Electrode Resistance
54
 The RL Drive Amplifier
RL Drive Simulation Showing Instability in the RLD
Feedback Loop
T 120.00

100.00
> 20dB/dec ROC (Rate of
80.00 Closure) = INSTABILITY
AOL
Gain (dB)

60.00

40.00
1/Beta
20.00

0.00

-20.00
1.00 10.00 100.00 1.00k 10.00k 100.00k 1.00M 10.00M
Frequency (Hz)
55
 The RL Drive Amplifier
Using RLD Simulation to Compensate for 1/Beta
Variation With Electrode Resistance
T 120.00

100.00
Intersection
80.00 of 1/ß and
AOL curve > -
40dB/dec
Gain (dB)

60.00

40.00
Feedback Comp
Placed vs. Worst
20.00
Case Electrode
Resistance and
0.00 RLD AOL

-20.00
1.00 10.00 100.00 1.00k 10.00k 100.00k 1.00M 10.00M
Frequency (Hz)
56
 The RL Drive Amplifier
RL Drive Stability Simulation Circuit of Feedback #1

VCC

Vref
+
+
U1 OPA333

R18 10M
R15 1M -
C2 47n
VCC
R8 63.4k R9 63.4k
R4 52k C10 1u

R12 1M
V3 0

C5 33p
C6 33p
+
V+
RG

C9 330p
RG1 25k Ref
V4 0C4 47n Out

C8 33p

R17 1M
RG
V-

C7 33p
RG2 25k - U3 INA333
R7 52k
R_RLD 52k

R10 63.4k R11 63.4k

C_RLD 47n

R19 10M VCC

L1 1G R3 10k

R2 100k -
Vref V1 2.5
Vloop -
+
+
U4 OPA333
VoA
R13 100k

+
Vref +
R14 100k
C1 1G

VCC
U2 OPA333 V2 2.5

R16 10M
VCC
+

Vin

57
 The RL Drive Amplifier
RL Drive Stability Simulation Circuit of Feedback #2

VCC

Vref
+
+

U1 OPA333
R15 1M -
C2 47n
VCC
R8 63.4k R9 63.4k
V3 0 R4 52k C10 1u

R12 1M
C5 33p
C6 33p
+
V+
RG

C9 330p
RG1 25k Ref
C4 47n Out

C8 33p
V4 0

R17 1M
RG
V-

C7 33p
RG2 25k - U3 INA333
R7 52k
R_RLD 52k

R10 63.4k R11 63.4k

C_RLD 47n

R1 10M VCC
Vref
L1 1G R3 10k
R2 100k -
Vref V1 2.5
-
Vloop
+
+ U4 OPA333
VoA +
R13 100k

+
Vref R14 100k
C1 1G

VCC V2 2.5

R16 10M
U2 OPA333 VCC
+

Vin

58
 The RL Drive Amplifier
RLD Stability Circuit with Compensated Amplifier

VCC

Vref
+
+

U1 OPA333
R15 1M -
C2 47n
VCC
R8 63.4k R9 63.4k
R4 52k C10 1u

R12 1M
V3 0

C5 33p
C6 33p
+
V+
RG

C9 330p
Vout
+

RG1 25k Ref

Vcm V4 0C4 47n Out

C8 33p

R17 1M
RG
V-

C7 33p
RG2 25k - U3 INA333
R7 52k
R_RLD 52k

R10 63.4k R11 63.4k

C_RLD 47n

VCC
R5 10M
NS1 2.5
R2 100k C3 3.3n NS
R1 10k Vref V1 2.5
-
U4 OPA333
R3 10k +
+

R14 100k
-
V2 2.5

R16 10M
+
VCC
+
R13 100k

Vref
VCC
U2 OPA333
59
 The RL Drive Amplifier
RL Drive Stability Simulation of Separate Feedback
Paths
T 120.00

AOL INA+ Electrode

100.00 RLD Local R Loop Feedback
Feedback
80.00
Gain (dB)

60.00

40.00

20.00 Compensation Composite

Feedback Feedback
0.00 1/ß

-20.00
1.00 10.00 100.00 1.00k 10.00k 100.00k 1.00M 10.00M
Frequency (Hz)

60
 The RL Drive Amplifier
Compensated RLD Circuit Simulation of 1/Beta and
AOL Intersection

T 120.00

100.00

Intersection of 1/Beta and

80.00
AOL curve is 20dB/dec =
STABLE
Gain (dB)

60.00

40.00

20.00

0.00

-20.00
1.00 3.16k 10.00M
Frequency (Hz)
61
 The RL Drive Amplifier
Gain and Phase Margin Plots of Compensated RLD
Amplifier

T 120.00
100.00

80.00
Gain (dB)

60.00
Loop Gain Phase Margin
40.00 @ 0dB = 70 degrees
20.00

0.00
180.00

135.00
Phase [deg]

90.00

45.00

0.00
1.00 3.16k 10.00M
Frequency (Hz)
62
 The RL Drive Amplifier
Step Response of RLD Amplifier and ECG Output

T 100.00m

Vin

-100.00m
2.55

VoA

2.45
2.80

Vout

2.20
0.00 5.00m 10.00m
Time (s)

63
The ECG Shield Drive

641
 The ECG Shield Drive
Shield drive eliminates leakage to ECG Inputs

ECGP CP1

–
ECGN CP2
VCC/2
+ + + • Shield is driven to
– (VIN(+) — VIN(–)) /2
– – • Eliminates Leakage
from CP1 and CP2

• Capacitance of cable can be 500 pF to 1.5 nF

• Isolation resistor Necessary for improved EMI/RFI filtering 65
 The ECG Shield Drive
AC Stability Simulation Circuit for OPA333 as Shield Driver

VCC
NS3 2.5
NS

Vref
+
+
U1 OPA333
R15 1M -

VF2 VCC
R4 10k C2 47n R8 63.4k R9 63.4k
C10 1u

R12 1M
C5 33p
C6 33p
R5 1k +
V+
RG

C9 330p
Vref R1 10k Ref VF1
Out
C8 33p
C4 47n
RG
V-

C7 33p
R2 10k -
R7 1k U2 INA333
R3 10k R10 63.4k R11 63.4k

VF3

VCC
VinP

VinN
U3 OPA333
+
+

Effective -
L1 1T VF

Cable Shield VG1

C1 1n

C3 1T
Capacitance Vout
+

66
 The ECG Shield Drive
AOL + 1/Beta Response of OPA333 Shield Drive and 1nF
Cable Capacitance

120.00

100.00

80.00
Gain (dB)

60.00
-40dB/decade Intersection
of AOL and 1/Beta =
40.00
UNSTABLE
20.00

0.00

-20.00
1.00 10.00 100.00 1.00k 10.00k 100.00k 1.00M
Frequency (Hz)
67
 The ECG Shield Drive
TINA Simulation Circuit for Stabilized OPA333 Shield Driver

VCC
NS3 2.5
NS

Vref
+
+

U1 OPA333
R15 1M -

VF2 VCC
R4 10k C2 47n R8 63.4k R9 63.4k
C10 1u

R12 1M
C5 33p
C6 33p
R5 1k +
V+
RG

C9 330p
Vref R1 10k Ref VF1
Out
C4 47n C8 33p
RG
V-

C7 33p
R2 10k -
R7 1k U2 INA333
R3 10k R10 63.4k R11 63.4k

VF3

VCC
VinP

U3 OPA333 C3 1T
R6 2k +
+
L1 1T VF
-
VG1

+
C1 1n

Vout
Shield Drive C11 1n

Compensation
Network R13 10k
68
 The ECG Shield Drive
TINA Simulation Shows > 45 Degrees Phase Margin for
OPA333 Shield Driver
120.00

100.00

80.00
Gain (dB)

60.00

40.00

20.00

0.00
0.00

-45.00
Phase [deg]

-90.00

-135.00

-180.00
1.00 1.00k 1.00M
Frequency (Hz)
69
Lead Off Detection

701
 Lead Off Detection
Lead Off Differentiates a Bad Lead from an Arrythmia

Body-Electrode VTH
VCC
Model

VREF
VTH

•Pull up Resistors Force +IN to Comparator High When Lead is Removed

•Comparator Voltage triggers ALERT
•Lead Off Indicative of ―Weak Lead‖
71
 Lead Off Detection
TINA Simulation Circuit for Lead Off Detect
VCC

Voltage-Controlled

R13 10M
VCC
Switches Simulate VR
VF3
U1 TLV3701
Disconnected Lead +
+

R14 13.5M
VS
VCC

R6 1M
- +

R15 10M
VM1

+
VF1

-
R1 1k C2 47n -

SW1 1
VCC Vref
R8 63.4k R9 63.4k

C5 33p
R4 52k

C6 33p
+
NS1 2.5 V+
RG

C9 330p
NS RG1 10k Ref Vout
Out

C8 33p
RG
V- VS

C7 33p
C4 47n
RG2 10k -

R12 1M
U3 INA333
R10 63.4k R11 63.4k

+
SW2 1

R3 10M
R7 52k VF2
R2 1k VG1
+

VS2 VCC

C1 10n U2 TLV3701 VS2

C3 47n R22 10k -
VR

+
R5 1M
R17 1M +
+
+ VG2
R21 52k VM2
R18 100k -
R16 10k - VCC
U5 OPA348
- U4 OPA348
+
R19 1M

+
+
R20 1M

VCC
VCC Vref
72
 Lead Off Detection
TINA Simulation Results for Lead Off Detect

T 2.00

Vswitch

0.00
4.99

VComparator

0.00
401.55u 559.69u 717.83u
Time (s)
73
Pace Detection

741
 Pace Detect
Pace Maker Pulse Specifications
dp

ap
t0

ao

ap = Amplitude (2-700mV)
ao = Overshoot
dp = Pulse Width (.1-100us)
t0 = Overshoot Time Constant (4-100ms)
Rise Time = 100us 75
 Pace Detect
Pace Detect Circuitry in Parallel with ECG Signal Path

VCC

C2 100p

R5 2M
Vpace Pos VCC
47n
VCC Vref
+ R4 100k
63.4k VECG_block +
TLV3701
+

+
ECGp

R6 2M
52k
+
33p

V+ Vout
- 10n R2 1k -
+
10k
RG
C9 330p

Ref +
Vpace1
Out + - -
+

OPA348

R1 100k
47n OPA348 VPDetect
ECGn RG

R17 1M
V- +
VCC +
10k

R3 1M
- VCC
33p

U5 INA333 -
Vpace Neg 63.4k
VCC
52k
100k

VCC Vpace2

R7 2M
+ +
+
Vref +
10n
- TLV3701
•AC Coupled Input Blocks ECG

R8 400k
R20 100k

Signal and Retains the PACER

U7 OPA348 R18 10k -
Pulse
-
+ •Window Comparator Triggers if
R22 1M

+
+
PACER Signal Detected
R23 1M

+ U6 OPA348
VCC
VCC Vref •Separate PACER Processing
Circuitry

76
 Pace Detect
PACE Signal Extracted From PACE + ECG Waveform

2.51

Vout

2.50
1.00m

Vpace Pos

0.00
153.93m 209.78m 265.62m
Time (s)

77
INA Post Gain and Filtering

781
 INA Post Gain and Filtering
Choice of High Gain + SAR ADC OR Low Gain + 24 bit Delta
Sigma ADC

At
At HighGain
High Gain At
At At
At Lowgain
Low gain At
At
electrode
electrode withlow
with lownoise
noise ADC
ADC electrode
electrode ADC
ADC
amplifiers
amplifiers

x200 x5
Amplitude

Amplitude
Amplitude

Amplitude
Noisefree
Noise free
Dynamicrange
Dynamic range

Noise
Noisefree
free
Dynamic
Dynamicrange
range
ADC
ADC
Amp
Amp Noise
Noise
noise
noise

Signal
SignalChain
Chain Signal
SignalChain
Chain

a)
a)Using
Usingaalow
lowresolution
resolutionADC
ADC b)
b)Using
Usingaahigh
highresolution
resolutionADC
ADC

SAR + filter Option Results in Same Input-Referred Noise as

the DC Coupled Delta-Sigma, but at what COST?

79
 INA Post Gain and Filtering
INA + Post Gain Amp With Differential Noise Source
R21 2M

VCC
Noise NS5 2.5
+ NS

Vref
C1 10p
R1 1k C2 47n +
+

U1 OPA333
R22 1M -
R18 2M
R4 1k VCC
R15 2M

R8 63.4k R9 63.4k R19 10k

+

ECGp
C10 1u

C5 33p

R23 1M
C6 33p
+ VF1
-
NS4 2.5

V+ Vref
NS

RG U5 OPA333
C9 330p RG1 10k Ref Vout
+
Out +
+

C8 33p

ECGn RG
R20 2M

V-

C7 33p

R17 10M
RG2 10k - NS VCC
U3 INA333
R10 63.4k R11 63.4k
NS2 2.5
C4 47n

R12 1M
R2 1k R7 1k

R5 1M
NS1 2.5
NS
R6 1k R3 10k
- -
U2 OPA333 U4 OPA333
+ +
+ Vref +
R14 1M
R13 1M

R16 1M

VCC VCC

80
 INA Post Gain and Filtering
Noise Coupled Differentially Translates to Output

1.00

50/60Hz

-1.00
4.33

50/60 Hz Couples Differentially;

Vout
How Do We Get Rid of It?

2.16
0.00 187.93m 375.86m
Time (s)
81
INA Post Gain and Filtering
Use Filter Pro to Design a 50/60 Hz Notch

82
 INA Post Gain and Filtering
ECG Circuit with Added 50/60Hz Notch + Post Gain

VCC
U3 OPA333
Vref
R21 2M
+
+
R1 1k C2 47n Noise
+ -

C1 10p Vref
R19 10k R18 1M
R4 52k VCC C11 1u
VCC
R15 2M

R8 63.4k R9 63.4k
+

ECGp

R30 1M
U4 OPA348 -
VF1

C5 33p
C6 33p

+ Vout + U7 OPA378
NS4 2.5

V+ +
NS

RG C12 330n +
+
C9 330p

RG1 10k Ref

VF2
Out -
+

C8 33p

R29 8.06k

R17 1M
ECGn RG R26 8.06k VCC
R20 2M

C7 33p
V- R24 64.9k
RG2 10k - U5 INA333_C
C3 330n

R22 1M
R10 63.4k R11 63.4k

R25 64.9k
C4 47n

R12 1M
NS1 2.5 -
R2 1k R7 52k C10 1n
NS U1 OPA348
R3 10k Vref NS
+
+
-

R27 8.06k
R5 1M
NS3 2.5
U6 OPA348
+ VCC
+
R14 1M

R6 1k
- R28 8.06k
R16 1M

VCC

R23 1M
U2 OPA348
+
+ Vref
R13 1M

VCC

83
 INA Post Gain and Filtering
Plots of ECG Output with Gain and 60Hz Notch
968.42m

Noise

-1.00
3.40

Vout_Filter

2.37
2.51

Vout_PostGain

2.50
2.51

Vout_INA

2.50

94.83m 237.76m 380.69m

Time (s)
84
 INA Post Gain and Filtering
Line Cycle Sampling with SAR converter on ‘T’ Wave at
Common Frequency Multiples of 50/60Hz
494.51m

370.88m
ECG (V)

247.25m

123.63m

0.00
400m 417m 433m 450m 467m 483m 500m 517m 533m 550m
Time (s)

Fsample (Hz) = n*(1/50Hz + 1/60Hz)-1 = n*(27.27) Hz

85
 INA Post Gain and Filtering
Comparison of Delta Sigma ADC vs. Lower
Resolution SAR ADC
Using a low resolution ADC00
x5 0.05 Hz x32 100 Hz
Elec 1
INA
Elec 2

DC blocking Additional High order

HPF gain Anti aliasing DOUT
Patient
filter
Protection, MUX ADC
Lead
selection
16 bit,

Elec 8
100KSps
` •Reduced Hardware
•Filter Requirements Relaxed
INA

Elec 9

•Lower Power
RL
•Lower System Cost
x5
Using a high resolution ADC
100 Hz
•Electrode Offset Info
Elec 1 ADS1258 Retained
INA
Elec 2

Patient
Protection,
Simple RC filter DOUT
MUX
Lead
selection
ADC

24 bit,
Elec 8 INA 100KSps
`

Elec 9

RL

86
 INA Post Gain and Filtering
Block Diagram of INA Gain, Simple RC Filter, and ADS1258

x5 x5 100 Hz 100 Hz
Elec 1 Elec 1 ADS1258 ADS1258
INA INA
lec 2 Elec 2

Patient Patient
Protection, Protection, DO
Simple RC filter
Simple RC filter MUX MUX DOUT
Lead Lead
selection selection
ADC ADC

24 bit, 24 bit,
lec 8 Elec 8 INA INA 100KSps
`
100KSps
`

Elec 9 Elec 9

RL RL

A single ADC in the MUX approach does not

necessarily mean lower power due to the higher
speed needed to perform MUX switching

87
ADS1298 Introduction

881
The ADS129x
The All-In-One ECG Chip

89
 ADS129x
Input Amplifier Specifications for Single Channel AFE
CH1P
+
PGA +
- BUF
-

RLD 24-bit
Delta-Sigma
ADC

+
+
BUF
-
PGA
CH1M -

•Low Input Voltage Noise

CMOS input PGA •Current Noise = 0.1pA√Hz Over Bandwidth
High input impedance •No Output Phase Reversal
•100dB AOL @ 50/60 Hz
Low input current noise •90dB CMR @ 50/60 Hz
Rail to rail input •IB = 150pA MAX vs. Temperature
•ZIN >100MΩ, CIN = 100pF max
90
 ADS129x
Noise Model for ECG AFE

• Noise is optimized with amplifier gain=4

• The 4uV p-p includes the crest factor of 6.6 to convert rms to pk-pk
• Noise is referred to the input

CH1P AFE

CH1M +
ADC

91
 ADS129x
Programmable Data Rates for Low Power and High
Resolution Modes

92
 ADS129x
MUX Selects Inputs to Front End PGA

• Normal Electrode
• Input Shorted
• RLD Input
• VDD
• TMP Sensor
• Input Test Signal

93
 ADS129x
Wilson Central Terminal

Wilson Central
Voltage

0.333(RA+LA+LL)

Augmented Lead
Voltages
0.5(RA+LA)

0.5(LA+LL)

0.5(RA+LL)

*The Same Amplifiers Used to Derive the WCT Voltage Can be

Switched to Obtain the Augmented Leads 94
 ADS129x
Input Amplifier and RLD Selection

VO = VS(1+R2/R1) + VCM1 + (VCM2 – VCM1) x (R2/R1)

Lead I configuration +/-VCM2 + VS AVDD

+ VO1
RS
-

Right leg drive

R1 R2 200K
External R-C

AVDD RLD Sum

5M (Register set)
ADC2
100K - 200K
RS +

VCM1 = AVDD/2

+/-VCM2 + VS AVDD
+ VO2
RS
-

R1 R2

*Compensation of RLD Amplifier is Based on the Gain Selected

and the Number of Amplifiers in the Loop
95
 ADS129x
RLD Selection—8 Channel Case
CH1P + 220K
RLDP1 + CH2N
- 50K
220K
RLDP2
ADS129x
20K 50K - RLD_INV 220K
20K CH1P
+ 50K
220K +
50K
CH1N
- RLDM1 220K
RLDM2
- REXT CEXT
220K
CH8M
CH2N 2M 264pF
CH3P + 220K -
RLIM RLD_OUT
- 50K
RLDP3
220K + CH4P RLD
Electrode
RLD
+
AMP
RLDP4 AVDD - AVSS
20K 50K - RLDREF_INT 2

50K 20K
+ 220K
+
RLDM3 50K
CH3N
- 220K
-
R1
RLDM4
CH4N
CH5P + 220K
RLD_REF

RLDP5 /RLDREF_INT
- 50K 220K + CH6P

RLDP6
20K
50K - R2
50K 20K
+ 220K +
RLDM5 50K
CH5N
- 220K
-
RLDM6 CH6N
CH7P + 220K
RLDP7
- 50K 220K + CH8P

RLDP8
20K
50K - All switches can be controlled
50K 20K
+ 220K
+ by register settings
RLDM7 50K
CH7N
- 220K
-
RLDINV RLDM8
CH8N

Cext
Rext
10M
- *Multiple PGA Outputs Can Be Switched to
264pF RLDREF_INT
RLD
RLDOUT
AMP
+
AVDD - AVSS
Derive Inverting RLD
RLDREF_INT WCT_TO_RLD 2
RLDREF WCT
96
 ADS129x
RLD with Multiple Devices

To Input Mux

To Input Mux
Device 1 Device 2 Device 1
To Input Mux

Power Down

+ VA1-8 + VA1-8 + VA1-8

- - -
RLD RLD RLD RLD RLD RLD RLD RLD RLD RLD RLD RLD
IN REF OUT INV IN REF OUT INV IN REF OUT INV
To Input Mux

Device n Device 2 Device 1

To Input Mux

To Input Mux
Power Down Power Down

+ VA1-8 + VA1-8 + VA1-8

- - -
RLD RLD RLD RLD RLD RLD RLD RLD RLD RLD RLD RLD
IN REF OUT INV IN REF OUT INV IN REF OUT INV

*With Multiple Devices the RLD Output Becomes the

Amplified Difference Between RLD REF and the
Summation of Multiple Lead Outputs
97
 ADS129x
Pace Detect

>From CH1P + 500K

PACE_ODD [4:1]
*Separate Pace Amplifiers Allow
+
- 50K
00xx 500K
xx00
From CH2P<
External Processing of Pace
20K 50K -
+ 50K 500K
20K
+ Signal; All Channels do Not Have
50K
>From CH1N
- 00xx 500K
xx00
-
From CH2N<
to Operate at Higher Data Rates
>From CH3P + 500K

- 50K
01xx
500K + From CH4P<
xx01
20K 50K -
50K 20K
+ 500K + PACE_OUT1 VREFCM ADS129x
01xx 50K
>From CH3N
- 500K
-

100K
xx01 250K
From CH4N<
>From CH5P + 500K CH1P
PDB_PACE_ODD
- 50K
10xx
500K + From CH6P< + 250K
xx10 CH7P
20K
50K - PACE
AMP - 250K
PACE_ODD[0:1]
50K 20K
+ 500K +
CH1M

10xx 50K
>From CH5N
- 500K
-
100K 250K
CH7M
xx10 From CH6N<
>From CH7P + 500K
PACE_OUT2 VREFCM
11xx
- 50K 500K + From CH8P<

100K
xx11
20K
50K - 250K
CH2P
50K 20K PDB_PACE_EVEN
+ 500K + + 250K
50K
>From CH7N
- 11xx
VREFCM
500K
- PACE
CH8P
PACE_EVEN[0:1]
xx11 AMP - 250K
From CH8N<
100K

CH2M

PDB_PACE + 100K 250K

PACE GPIO1 PACE_IN(GPIO1) CH8M
TESTN_PACE_OUT2 AMP
-
(AVDD-AVSS)/2
100K PACE ODD PACE EVEN CHANNEL
200K

PACEO1 PACEO0 PACEE1 PACEE0 SELECTED

PDB_PACE + x x 0 0 Channel #1
0 0 x x Channel #2
TESTP_PACE_OUT1 PACE -
AMP x x 0 1 Channel #3
0 1 x x Channel #4
200K
x x 1 0 Channel #5
1 0 x x Channel #6
x x 1 1 Channel #7
1 1 x x Channel #8

98
 ADS129x
Lead Off Detection
AVDD AVSS

LOFF_FLIP
FLEAD_OFF[0:1]
FREQ = DC, FDR/2 or FDR/4

FLEAD_OFF[[1:0] VX
Patient Skin – Electrode Patient
Contact Model Protection Anti-Aliasing Filter

>
Resistor <512KHz

< R1 ~ 10M
Z1
LOFF_STATP
LOFF_SENSP LOFF_SENSM
47nF
& VLEAD_OFF_EN & VLEAD_OFF_EN

R2 VINP
51K
+
100K
EMI
Filter
To ADC
51K R2 VINM
100K
-
47nF LOFF_SENSP LOFF_SENSM
Z2 & VLEAD_OFF_EN & VLEAD_OFF_EN LOFF_STATM

LOFF_FLIP

4-bit COMP_TH[2:0]
25, 17.275, 12.5, & DAC
6.25nA
Z3 AVSS
AVDD
ILEAD_OFF[0:1]
51K R3
RLD OUT
100K 99
 ADS129x
Respiration Testing Measures the Change in Thoracic
Impedance with Inhalation of O2

ADS129x ADS129x
PATIENT’
PATIENT’ S CHEST
S CHEST RPROTEC RPROTEC
T E2 T
E4
100K 100K

RTHORAX
V
RPROTEC 64KHz
RPROTEC E1 T

T 100K
E3
100K
RTHORAX

V *AC Current is injected into the

64KHz
RPROTEC
T
Patient’s Thorax and the Change
E2
100K in Voltage is Measured to
Calculate Change in Impedance

RPROTEC
T
E1
100K

100
 ADS129x
Respiration Functions

ADS129x

GPIO1

GPIO
GPIO2

Changing Phase Allows

GPIO3/
Measurement/Compensation for

RESPIRATION
RESP_OUT

RESP_MODE
Complex Impedance Phase Shifts
Between Modulator and
GPIO4/
RESP_PH_OUT Demodulator
RESP_PH [4:2]

RESP_PH[4:2] bit CODE

PHASE
RESP_ RESP_ RESP_P SELECTED
PH2 PH1 H0
0 0 0 22.5 Deg
0 0 1 45 Deg
0 1 0 67.5 Deg
0 1 1 90 Deg
1 0 0 112.5 Deg
1 0 1 135 Deg
1 1 0 157.5 Deg
1 1 1 180 Deg

101
 ADS129x
Internal Voltage Reference

Simplified ADS129x internal reference block diagram

The Internal Band Gap Accuracy = 1%

Internal REF can be Powered Down
VREFP Can Be Supplied Externally
102
Thank You
Contact Information:
hann_matthew@ti.com

103
Questions?

1041
 Acknowledgements
• Beraducci, Mark and Soundarapandian, Karthik. Sbaa160, Application Report: Analog
Front End Design of ECG Systems Using Delta-Sigma ADCs. March 2009.
• Brown, John --Burr Brown Strategic Marketer, general information
• Brown, John and Joseph Carr. Introduction to Biomedical Equipment Technology.
Prentice Hall Inc. New Jersey. 1981, 1993.
• Dubin, Dale. Rapid Interpretation of EKG’s. Cover Publishing Company. Fort Myers.
2000.
• Fraden, Jacob. Handbook of Modern Sensors—Physics, Designs, and Applications.
Advanced Monitors Corporation. San Diego. 2004.
• Franco, Sergio. Design with Operational Amplifiers. McGraw-Hill Inc. NY. 1988.
• Fredericksen, Thomas M. Intuitive Operational Amplifiers. McGraw-Hill Inc. 1988.
• Graeme, Toby, Huelsman. Operational Amplifiers—Design and Applications. McGraw-
Hill Publishing Company. New York. 1971.
• Gray, Paul R. and Meyer, Robert G. Analysis and Design of Analog Integrated Circuits.
John Wiley & Sons. New York. 1977
• Green, Timothy - Linear Applications Manager, general information
• Kuehl, Thomas--Linear Applications Engineer, general information
• Norton, Harry N. Sensor and Analyzer Handbook. Prentice Hall Inc. New Jersey. 1982.
• Soundarapandian, Karthik--Over sampling Manager, general information

105